W3C home > Mailing lists > Public > public-ws-addressing@w3.org > August 2006

Re: Proposal for CR33

From: Alastair Green <alastair.green@choreology.com>
Date: Wed, 16 Aug 2006 15:58:36 +0100
Message-ID: <44E3329C.2040604@choreology.com>
To: Anish Karmarkar <Anish.Karmarkar@oracle.com>
CC: "public-ws-addressing@w3.org" <public-ws-addressing@w3.org>, Doug Davis <dug@us.ibm.com>

Anish, Doug --

This seems like an elegant solution. I like the shift in focus from 
prescription to permission. This presumably includes the ability to use 
[reply endpoint] = none, which is a non-addressable URI (I think this is 
CR32).

What happened to the faults to allow reaction by a conformant service to 
a non-conformant client? I assume this has been discussed already.

Can I confirm my understanding of how this would play when using RM 
MakeConnection?

If the contract is: service will accept RM anon URI, and will accept 
MakeConnections, then this could be expressed by a WSDL interface 
containing 1..n endpoints that support back-channel responses, plus an 
endpoint for MakeConnection that mandates a back-channel response. This 
would (roughly) signal to the client that the service is able to 
converse via RM polling.

If RM wanted to go further, it could add a WSDL or policy declaration 
that means "I understand WS-RM anon URI; the sender can get messages 
that ensue on this operation, via MakeConnection". That is outwith 
WS-A's concerns.

* * *

I'm left with one question, which I've raised before. Does the use of 
WS-A anon as the value for the response endpoints, permit an empty (no 
SOAP envelope) response (just an ack)?

The answer to this question does not directly affect the proposed 
solution to CR33, but it is relevant to understanding whether WS-RM was 
compelled to invent its own anonymous URI.

Another way of looking at the question, from a WSDL angle: is [reply 
endpoint] = WS-A anon compatible with WSDL 2.0 MEP In-optional-out 
(which is the natural MEP for the MakeConnection use-case)? In the WS-A 
SOAP Binding I see at least a strong implication that the standard WS-A 
anon URI precludes an empty response.

Yours,

Alastair




Anish Karmarkar wrote:
> Dug and I took an action during this week's call to send out a 
> proposal for CR33 [1]. Word version of the proposal is attached. It is 
> a marked up version of section 3.2 [2] of the WS-Addressing 1.0 -- 
> WSDL Binding spec. PDF (diffed and non-diffed  versions) are also 
> attached.
>
> -Anish
> -- 
>
> [1] http://www.w3.org/2002/ws/addr/cr-issues/Overview.html#cr33
> [2] http://www.w3.org/TR/2006/CR-ws-addr-wsdl-20060529/#anonelement
Received on Wednesday, 16 August 2006 14:58:58 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:35:14 GMT